342 OSCAR RIDDLE. 



conceivable almost a necessity that each variety should at 

 least sometimes be able to force its stable character (if these do 

 not both rest directly upon the same process, e. g., ty rosin oxi- 

 dation) practically unchanged into a new combination and pro- 

 duce a new form --an animal with the color of one parent and 

 the size of the other. 



This clearly brings us face to face with something resembling 

 unit characters and particulate inheritance. I can see no reason, 

 from studies on the nature of color formation, and from the 

 necessary deductions as to the way in which color must be trans- 

 mitted, to doubt the possibility or the probability of the formation 

 of races with new characters, or rather a race with a new combi- 

 nation of old characters ; and with some such newly combined 

 characters in very stable equilibrium, i. e., breeding true. In 

 fact, it is the recognition of the state in which our color character 

 exists in the germ, that is, as a given pitch of power for oxidizing 

 tyrosin compounds, and this not latent but active in the germ as 

 later, that enables us to view the mechanism by which such a 

 result is brought about. 



The demonstration of the existence of such combinations of 

 characters is, I believe, the supreme contribution of Mendelism to 

 our knowledge of heredity. Phenomena of dominance, of segre- 

 gation (?) and proportion, are but minor and special manifesta- 

 tions of a process much more important, general and inclusive ; 

 and which general process is, in color inheritance at any rate, 

 the propagation and occasional quantitative modification (four 

 types above) of oxidizing powers, with their more or less con- 

 stancy of expression through settling into points of easy or fixed 

 oxidizing equilibrium. 



Our view does not, however, allow the acceptance of the unit 

 character hypothesis without very considerable and rather radical 

 modification. The prevailing idea among Mendelian workers 

 has been essentially that each character, each recognizable differ- 

 entiation, each member of a group of factors that forms a char- 

 acter, is quite separate and capable of being shuffled in the germs, 

 and of independent appearance in the zygote. Now, as already 

 noted, experience with the melanins drives home the point that 

 a long series of very distinct characters have not each even one 



